Another CPU release from Intel? 

It hasn’t even been two weeks since Intel “released” their 1GHz Pentium III in limited quantities to select OEMs and today they’re back with yet another processor launch.  More than a few members of the industry have become disgusted with the current state of affairs in the CPU industry, especially with the clock speed battle going on between Intel and AMD. 

If you recall, Intel’s 1GHz Pentium III announcement was in direct response to AMD’s 1GHz Athlon announcement just two days earlier.  Need proof that it was in direct response to that announcement?  Because unlike AMD, who was launching 900 and 950MHz Athlon parts alongside their 1GHz chips, Intel jumped straight from the Pentium III 800EB to the new 1GHz parts leaving a huge gap in clock speed, something that is completely irregular for Intel.  

The 1GHz Pentium III was launched earlier than expected, but in spite of that, Intel chose to stick to their release schedule for the rest of their Pentium III line because those processors, unlike the 1GHz CPUs, were to be made available to the general public, not only Dell, HP and IBM. 

In order to fill the gap between the Pentium III 800EB and the Pentium III 1GHz, Intel is continuing with the release of their 850 and 866MHz versions of the Pentium III processor. 

For the first time since Q3-99, Intel will be releasing a Pentium III CPU without pushing AMD to release a competing product.  Why?  Because AMD has already announced their 850 – 950MHz parts, thus completing their Athlon line with a total of 11 processors running at 500MHz all the way up to 1GHz in 50MHz steps. 

In essence, Intel is playing catch-up with these two parts being announced today.

Pentium III 866 Specifications

·        29 million transistor 0.18-micron Coppermine core

·        866MHz clock speed – 6.5x clock multiplier

·        32KB on-die L1 cache running at core speed

·        256-bit Advanced Transfer Cache - 256KB on-die L2 cache running at core speed

·        Advanced System Buffering

·        242-pin Slot-1 GTL+ CPU interface running at 133MHz

·        1.65v core voltage

The 833MHz Pentium III is not any different from the rest of the Pentium III line; it is based on the 0.18-micron Coppermine core and actually uses the same stepping (cA2) of the core as the previous Pentium IIIs with the exception of the 1GHz Pentium III, which is using a new core stepping required to hit the magical 1000MHz clock speed. 

The 866MHz chips are still using the 1.65v core voltage that is standard for all Pentium III processors with the exception of the lone 1GHz ranger that uses an increased 1.70v to, once again, hit the 1GHz clock speed. 

If you’re not familiar with some of the features the Coppermine core offers, the two biggest and most talked about benefits of this core are the Advanced Transfer Cache (ATC) and the Advanced System Buffering (ASB). 

The ATC on the Pentium III is just the fancy name for the on-die 256KB L2 cache.  Keep in mind that the Pentium III isn’t just a Celeron with twice as much cache and SSE, the L2 cache bus has been increased from the 64-bit bus width on the older Pentium III and Celeron processors to 256-bits wide.  The ATC also refers to the 8-way associativity of the 256KB L2 cache on the Coppermine (compared to the 4-way associative L2 cache on the old Pentium III/Celeron). 

Because the 256KB of L2 cache is on-die, the transistor count of the Pentium III is increased tremendously over the 9.5 million transistors that made up the original Pentium III (Katmai) core.  The 256KB L2 adds about 19 to 20 million transistors, putting the total transistor count of the Coppermine at approximately 29 million transistors. 

Advanced System Buffering is a simple term that represents the increase in buffers the Pentium III Coppermine offers over the previous generation of processors, including the Katmai based Pentium IIIs.  If you are interested in specifics, there are now 6 fill buffers, 8 bus queue entries, and 4 writeback buffers (up from 4, 4, and 1 respectively).  These three optimizations all help to take advantage of the 1.06GB/s bandwidth offered by the 133MHz FSB. 

Pentium III 850 Specifications

·        29 million transistor 0.18-micron Coppermine core

·        850MHz clock speed – 8.5x clock multiplier

·        32KB on-die L1 cache running at core speed

·        256-bit Advanced Transfer Cache - 256KB on-die L2 cache running at core speed

·        Advanced System Buffering

·        242-pin Slot-1 GTL+ CPU interface running at 100MHz

·        1.65v core voltage

The Pentium III 850 is identical to the other two chips with the exception of its specified FSB frequency, which is 100MHz instead of the 133MHz for the 866 chip. 

This makes the Pentium III 850 the ideal solution for your BX boards if you still haven’t migrated to a 133MHz FSB motherboard platform which, judging by the incredible cost of RDRAM and the just recent availability of good VIA 133A boards, we can’t really blame you for. 

Unfortunately, as of the time of publication, there were no BIOS updates for any of the boards we had in lab that would enable the 8.5x clock multiplier required by the Pentium III 850.  While this shouldn’t be a problem for most users since your CPU will feature a locked 8.5x multiplier, our test sample did not feature a locked multiplier and thus we were unable to run the 850MHz tests. 

We will provide an update as soon as we receive a BIOS update from the motherboard manufacturers that enables the 8.5x clock multiplier.

The many flavors of Coppermine

Before we dive into the new chips, let’s get a small update on what CPUs are now available in the Pentium III family:

FC-PGA is here

As we mentioned in our review of the first Pentium IIIs based on the Coppermine core, by the middle of this year, over half of all desktop Pentium III CPUs will be available in FC-PGA versions. 

At this point in time, you can get both 133MHz FSB and 100MHz FSB Pentium III CPUs in speeds up to 750MHz in FC-PGA Socket-370 versions.  The major reason for moving to FC-PGA Socket-370 is that it is cheaper for Intel to produce a socketed chip than it is for them to produce essentially the same chip but on a SECC2 card where most of the PCB space is wasted. 

From the prices we have seen, SECC2 Pentium IIIs (Slot-1) are running for anywhere from 1% up to 10% more expensive than their FC-PGA counterparts.  There is a definite shift towards FC-PGA, and by the end of this year, we can expect almost all desktop Pentium III processors to be FC-PGA Socket-370 chips. 

We are noticing a similar push by motherboard manufacturers towards their Socket-370 motherboards with Pentium III (FC-PGA) support.  In the coming few months, expect more Socket-370 motherboards to hit the streets, including dual Socket-370 (FC-PGA) motherboards. 

Currently FC-PGA Pentium IIIs will not work in dual processor mode using Socket-370 to Slot-1 converter cards with SMP support for Celerons.  The Dual Processor Celeron trick does not work with the FC-PGA Pentium IIIs, although it is very clear that Intel does have plans to take FC-PGA parts to the dual processor world as indicated by their chipset design guide entitled: Dual FC-PGA Intel Pentium III Processor and Intel 840 Chipset Design Guide.

For now, if you’re looking to run a dual Pentium III (Coppermine) system, you’re in luck, because every Pentium III CPU is now available in a SECC2 interface (including the 500E/550E) which do currently work in dual processor configurations.  The problem with FC-PGA CPUs right now is that no one has discovered how to get them properly working in dual processor mode with something like the Dual Celeron trick. 

We experimented with some possibilities in the lab last weekend and managed to get a dual FC-PGA system to boot using two modified Iwill Slocket-II adapters, but unfortunately, the setup was not stable enough to be considered a viable solution.  But, once again, since all Pentium III CPUs are available in SECC2, if you really want dual Pentium IIIs then you can simply get the Slot-1 versions that currently work properly in dual, otherwise wait for more results on what modifications need to be made to the current crop of Socket-370 to Slot-1 adapters. 

Availability and Pricing

The Pentium III is now much more available than it was just a couple months ago, especially in the higher clock speed parts.  While it is still much easier and much cheaper to pick up an Athlon, the Pentium III is currently available at speeds of up to 800MHz from on-line vendors. 

You can expect the 850 and 866MHz parts to pop up over the next few weeks, but for the most part, most of them to be bought by the big OEMs, such as Dell, HP, and IBM, which are currently the only ones that offer 1GHz systems based on the Pentium III processor.  So for most AnandTech readers, don’t expect to be able to find 850 or 866MHz Pentium IIIs available from too many online vendors.

At the same time, the 900/950MHz Athlons aren’t really available outside of the major OEMs, so as of now, you don’t really have an alternative if you want something faster than 850MHz.  But for most AnandTech readers, overclocking to the 900MHz+ range is even better than purchasing a chip originally clocked at 900MHz+.

Overclocking

As we mentioned earlier, the 1GHz Pentium III actually uses a new stepping of the Coppermine core; that stepping isn’t present on the newer 850 or 866MHz Pentium IIIs, and thus, we can’t really expect these chips to hit the 1GHz mark with much ease. 

Because of this, you shouldn’t expect the 850 or 866MHz parts to be great overclockers, at least not at the start.  As the yields on these chips improve, you can expect them to become better overclockers, but until then, you’re better off buying a slower processor and overclocking it to 800 – 900MHz. 

Motherboard Support

We mentioned, and you probably already know, that the i820 motherboard platform isn’t exactly the most popular platform in the market right now.  So what kind of support does the Pentium III have? 

Fortunately, VIA’s Apollo Pro 133A chipset is finally performing like it should and there are more than enough motherboards based on the chipset to provide you with a few good options when picking up a board. 

In particular, we used the ASUS P3V4X in our 133A tests with the 1GHz Pentium III simply because our Gigabyte 133A board failed to properly support the CPU without an updated BIOS that was, at the time of testing, unavailable.  If you do recall from our VIA Apollo Pro 133/133A Motherboard Roundup for February 2000, the ASUS P3V4X performed at the top of its class under all of the business and content creation tests, but it dropped down to last place in the gaming tests. 

Click to Enlarge

Apparently, there are issues with the P3V4X and the VIA AGP GART driver v4.00 that cause the gaming performance of the setup to drop considerably.  The only solution for now is to use v3.59 of the GART drivers from VIA, which completely corrects the performance issues.  We are still awaiting an explanation for this anomaly, which is present only with the P3V4X, from ASUS, and when we do receive an official statement, we will let you know. 

In order to illustrate the maximum performance of this setup, we used v3.59 of the GART drivers with the P3V4X test bed.

The Test

 

The performance of the Pentium III under CC Winstone 2000 is pretty much identical to the Athlon on a clock for clock basis. The main difference here comes in the form of the various chipset/memory combinations that can be used with the setups.

Under a multitasking environment which CC Winstone 2000 helps to simulate, the VIA 133A chipset seems to take a back seat to the i820 + RDRAM combination but not by a large amount at all. With RDRAM still going for 3 to 5 times the cost of PC133 SDRAM, the small performance hit here is definitely worth it.

Contrary to what Intel may say, the Apollo Pro 133A chipset is definitely the way to go for 133MHz FSB support with the Pentium III.

SYSMark 2000 very clearly favors the Pentium III over the Athlon as is evident by the change in standings from CC Winstone 2000 to SYSMark 2000.

Once again we have the VIA 133A chipset taking a back seat to the i820 + RDRAM setup but if you factor in the cost of the latter, the 133A + PC133 setup becomes all the more attractive of an option.

The gaming performance advantage is still in Intel's hands with the Pentium III but depending on what type of gamer you are, this resolution may not be representative of real world gaming performance, thus making the next test even more important.

At 1024 x 768 x 32-bit color our test bed's DDR GeForce is fillrate limited and thus there is no real difference between the performance of any of the 28 CPUs compared in the above chart. This just helps to illustrate the point that clock speed isn't always king.

Under UnrealTournament we have the Pentium III once again taking the lead away from the Athlon, but if you look at the performance differences it is obvious that the final decision should come down to the overall affordability/availability of the particular CPU in question.

You are not going to notice a 2 fps difference in performance making the Pentium III and Athlon equally viable options here for the avid gamer. If you're going for the Pentium III once again, the VIA 133A chipset is a much more intelligent option than the i820 + RDRAM combo simply because of cost.

There are obviously a number of other limitations in this test outside of the CPU which is explained by the relatively small difference in performance between the top 15 processors on this chart.

These limitations are mainly a result of the UT engine itself which explains why the performance doesn't change all that much when going from 640 x 480 to 1024 x 768 at least in comparison to a game like Quake III Arena where the difference is huge because of a clear limitation kicking in.

We get the same old picture under Expendable at 640 x 480, the Pentium III comes out on top of the Athlon by a small amount and the i820 + RDRAM platform comes out on top of the VIA 133A platform by little bit as well.

SPECviewperf

The Standard Performance Evaluation Corporation, commonly known as SPEC, managed to come up with a synthetic benchmark with real world implications. By running specific "viewsets" SPECviewperf can simulate performance under various applications. To be more accurate, according to SPEC, "A viewset is a group of individual runs of SPECviewperf that attempt to characterize the graphics rendering portion of an ISV's application." While this method is by no means capable of identifying the performance of a card in all situations, it does help to indicate the strengths and weaknesses of a particular setup.

SPECviewperf 6.1.1 currently features five viewsets: the Advanced Visualizer, the DesignReview, the Data Explorer, the Lightscape and the ProCDRS-02 viewset. Before each benchmark set we've provided SPEC's own description of that particular viewset so you can better understand what that particular viewset is measuring, performance-wise.

Each viewset is divided into a number of tests, ranging from 4 to 10 in quantity. These tests each stress a different performance element in the particular application that viewset is attempting to simulate. Since all applications focus on some features more than others, each one of these tests is weighted meaning that each test affects the final score differently, some more than others.

All results are reported in frames per second, so the higher the value, the better the performance is. The last result given for each of the viewsets is the WGM or Weighted Geometric Mean. This value is, as the name implies, the Weighted Geometric Mean of all of the test scores. The formula used to calculate the WGM is as follows:

With n being the number of tests in a viewset and w being the weight of each test expressed as a number between 0.0 and 1.0.

If you'd like to know more about why a Weighted Geometric Mean is used, SPEC has an excellent article detailing just why, here.

We ran the SPECviewperf 6.1.1 package under NT for a high-end workstation performance comparison. A noteworthy change has been made since the last Athlon CPU review, we have started using a DDR GeForce in the high-end tests after discovering that in spite of the GeForce's hardware T&L engine the card uses virtually all of the host CPU during the SPECviewperf tests thus putting a great strain on the CPU.

We also received an updated set of drivers (v3.76) for Windows NT that fully take advantage of the KX133 chipset which help to produce some very interesting benchmark results. An interesting thing to note is that because the VIA Apollo Pro 133A (VIA 133A) chipset uses the same AGP core as the KX133 chipset, the 3.76 drivers actually improved performance on the 133A platform as well.

Advanced Visualizer (AWadvs-03) Viewset

Taken from http://www.spec.org/gpc/opc.static/awadvs.htm

 
• 55% material shading (textured, z-buffered, backface-culled, 2 local lights)
• 95% perspective, 80% trilinear mipmapped, modulated (41.8%)
• 95% perspective, 20% nearest, modulated (10.45%)
• 5% ortho, 80% trilinear mipmapped, modulated (2.2%)
• 5% ortho, 20% nearest, modulated (.55%)
• 30% wireframe (no z-buffering, no lighting)
• 95% perspective (28.5%)
• 5% ortho (1.5%)
• 10% smooth shading (z-buffered, backface-culled, 2 local lights)
• 95% perspective (9.5%)
• 5% ortho (.5%)
• 5% flat shading (z-buffered, backface-culled, 2 local lights)
• 95% perspective (4.75%)
• 5% ortho (.25%)

The Awadvs test is obviously limited by the DDR GeForce used in our test bed, and thus processor speed doesn't affect the test to any noticeable degree.

The only outliers in this case are the AMD 750 tests with the Athlon and the Pentium III tests on the BX platform which both deliver noticeably less memory bandwidth (or in the case of the AMD 750 platform, exhibit a less efficient use of memory bandwidth) than the faster setups on the chart.

DesignReview (DRV-06) Viewset

Taken from http://www.spec.org/gpc/opc.static/drv.htm

The DRV-06 viewset is much more influenced by processor performance than the previous test which causes the nice performance curve we see above.

Just as the Pentium III held its ground in the gaming performance tests earlier, the Athlon is holding its ground here and quite well at that.

The 3.76 drivers from NVIDIA also seem to help the VIA 133A and KX133 chipsets improve performance here as they put the i820 + RDRAM setup to shame which brings up the question as to exactly what these drivers are doing that helps the VIA AGP core so well. It's obvious that these drivers aren't Athlon-only because their positive effects are felt by both the KX133 and 133A test platforms, so there's something that's shared by these two chipsets (their AGP core) that is performing much better with these drivers.

Whatever it is, professional users that are KX133 and 133A owners shouldn't be complaining at all but those that spent the $500 - $1000 on RDRAM have the right to feel a bit cheated.

Data Explorer (DX-05) Viewset

Taken from: http://www.spec.org/gpc/opc.static/dx.htm

The IBM Visualization Data Explorer (DX) is a general-purpose software package for scientific data visualization and analysis. It employs a data-flow driven client-server execution model and is currently available on Unix workstations from Silicon Graphics, IBM, Sun, Hewlett-Packard and Digital Equipment. The OpenGL port of Data Explorer was completed with the recent release of DX 2.1.

The tests visualize a set of particle traces through a vector flow field. The width of each tube represents the magnitude of the velocity vector at that location. Data such as this might result from simulations of fluid flow through a constriction. The object represented contains about 1,000 triangle meshes containing approximately 100 vertices each. This is a medium-sized data set for DX.

The DX-05 test, like the previous DRV-06 test shows off the power of the Athlon as well as the improvements provided by NVIDIA's 3.76 drivers on all VIA based platforms that use VIA's AGP 4X core (KX133 & 133A in this case).

Lightscape (Light-03) Viewset

Taken from: http://www.spec.org/gpc/opc.static/light.htm

The Lightscape Visualization System from Discreet Logic represents a new generation of computer graphics technology that combines proprietary radiosity algorithms with a physically based lighting interface.

There are four tests specified by the viewset that represent the most common operations performed by the Lightscape Visualization System

Nothing new here, the Athlon takes the lead, the VIA 133A + PC133 SDRAM comes out on top of the i820 + RDRAM setup, and the BX and AMD 750 platforms hold up the bottom end of the chart.

ProCDRS-02 Viewset

Taken from: http://www.spec.org/gpc/opc.static/procdrs.htm

The ProCDRS-02 viewset is a complete update of the CDRS-03 viewset. It is intended to model the graphics performance of Parametric Technology Corporation's CDRS industrial design software.

For more information on CDRS, see http://www.ptc.com/icem/products/cdrs/cdrs.htm

The viewset consists of ten tests, each of which represents a different mode of operation within CDRS. Two of the tests use a wireframe model, and the other tests use a shaded model. Each test returns a result in frames per second, and a composite score is calculated as a weighted geometric mean of the individual test results. The tests are weighted to represent the typical proportion of time a user would spend in each mode.

All tests run in display list mode. The wireframe tests use anti-aliased lines, since these are the default in CDRS. The shaded tests use one infinite light and two-sided lighting. The texture is a 512 by 512 pixel 24-bit color image.

The ProCDRS-02 test actually begins to show an advantage to RDRAM for the first time in the entire SPECviewperf suite. The performance advantage is definitely negligible at best, but it's there. The Athlon still comes out on top.

Conclusion

Just as we concluded in our Pentium III 1GHz and our Athlon 1GHz reviews the Pentium III holds the advantage under Win98 gaming performance while the Athlon takes the lead under professional level OpenGL performance under WinNT.

In terms of price, the advantage tilts to AMD as the Athlon (which is due for yet another price cut soon) is consistently priced lower and is more available than the Pentium III on a clock for clock basis. While the Pentium III is obviously much more available than it was back in February, it is still not nearly as abundant as the Athlon is, especially in the 700 - 800MHz clock speed range.

For overclockers, the best choice is to stick with one of the lower clock speed processors, either the Pentium III 500E/550E/600E or similarly clocked Athlons and push them to the higher clock speeds. In most cases, this is a very feasible option since the yields on these chips are high enough that the lower end CPUs can actually hit the higher end speeds without much trouble.

If you're considering one of these new Pentium IIIs, first of all, don't expect to be able to go online to your favorite vendor and find them with much ease. Secondly, Intel's next generation 0.18-micron Celerons are due out shortly, and if all goes according to plan these Celerons will be the next major overclockers since they will be based on the same 0.18-micron process as the current Coppermine Pentium IIIs. Since these Celerons will run on a 66MHz FSB by default, they will feature pretty high clock multipliers, luckily with the slowest chips most likely debuting in the 533 - 600MHz range we are expecting to be able to overclock these babies to 800 - 900MHz.

While we'll have to wait for the final word on that, considering the price point of the Celeron family of processors they are definitely worth waiting for to see if they do end up being the next big overclocker from Intel.

If you are intent on purchasing a Pentium III now, the best platform still seems to be the VIA 133A platform, especially with the actual availability of decent 133A based motherboards such as the ASUS P3V4X and the Gigabyte GA-6VX-4X.

For more information read Sharky Extreme's Pentium III 866 Review.

How it Rates

AnandTech CPU Rating
	Rating (x/10)
Performance Under content creation and gaming applications the Pentium III performs quite well, even outperforming the Athlon in some cases. Unfortunately the CPU loses points here because it doesn't perform nearly as well as the Athlon in the NT - OpenGL tests.	7.5
Features Features aren't a problem here. The full speed on-die L2 cache helps performance quite a bit, and in cases where SSE is taken advantage of the Pentium III definitely excels. One feature that is heavily in favor of the Pentium III is SMP motherboard and chipset support. While there will be a SMP Athlon chipset due out later this year/early next year, SMP support is something that professionals demand now, and the Pentium III delivers it.	7.5
Price The price to performance ratio of the Pentium III is less than that of the Athlon simply based on the recent price conditions where the Athlon is, clock for clock, noticeably cheaper than the Pentium III. Unless AMD or Intel experience a major change, don't expect things to turn around anytime soon.	6.0
Availability These new chips will be much more available than the 1GHz Pentium IIIs but the Pentium III in general is still not as widely available as the Athlon.	6.5
Overall Rating - Not an Average - Click here to find out why	7.0